What Makes Google Translate Better?

Google Translate is a beta service offered by Google designed to translate text into another language. Google uses its own state-of-the-art translation software, unlike other services that utilize SYSTRAN such as Babel Fish, Yahoo and AOL. It is based on an approach researched by Franz-Josef Och, called statistical machine translation that some suggest could result in a language industry revolution. Och now leads Google’s machine translation department.

Google Translate algorithms are based on statistical analysis as opposed to the more traditionally used rule-based analysis that defines vocabularies and grammars, and as such does not apply grammatical rules. It allows a reader to understand general context but does not promise accurate translation.

United Nations documents are the source of much of the vast amount of linguistic data used. With documents available in all six official UN languages – Chinese, English, Arabic, French, Spanish and Russian – Google has acquired 20 billion words to apply to translations. Google Translate is now able to offer translations between 34 languages.

Google Translate offers a “detect language” option to automatically identify the language of the text being translated. It’s accuracy is increased with more text entered. When a user sees “translate this page” next to a search result, clicking on the link will prompt Google to automatically translate the page into the user’s language.